The theory underlying our understanding of genetic differences between human populations relies implicitly on an assumption that emigration is "unstructured", i.e. that emigrant allele frequencies are distributed as though emigrants were drawn at random without replacement from their group of origin (Rogers and Harpending 1986). There is extensive evidence, however, that this assumption is often violated in nature, and theoretical studies have shown that such violations may have substantial effects on genetic differences between groups (Fix 1978; Leslie 1980; Rogers and Jorde in press; Rogers in press). The proposed research will estimate the magnitude of these violations in 24 populations and use the theoretical work just cited to assess their effect on genetic population differences.